It is well understood that customers recognize the enormous value of their personal data. The value exchange holds that customers will provide data to a brand in return for a personalized experience. In a 2022 Dynata customer experience (CX) survey, 59 percent of customers said that receiving personalized offers, discounts or perks was one of the top ways brands facilitate the sharing of more data. The willing exchange, however, comes to a screeching halt when brands share data with a third-party without consent. In the survey, nearly half (48 percent) of consumers said they will stop doing business with a brand that shared data without consent.
Brands, too, particularly retailers, media companies and walled garden behemoths (Amazon, Google) increasingly recognize both sides of the equation. That is, they recognize the value of their own first-party data and they see the untapped potential in monetizing that value, yet they insist on protecting consumer trust and honoring expectations surrounding data privacy.
Safeguard Privacy with a Data Clean Room
Enter data clean rooms, which are gaining traction as a way for companies to merge and match two or more first-party data sets without exposing any personally identifiable information (PII) that one company possesses about its customers to another party. Using a specific type of data encryption, data clean rooms allow a company to analyze, match and build models using anonymized data without ever accessing or decrypting PII.
Walmart Connect and Roku are prime examples of a retailer and media company monetizing brand data by using walled garden data clean rooms to sell digital advertising on branded sites – one of many data clean room use cases intended to enhance customer experience (CX). In this case, Walmart and Roku clients provide encrypted customer data and merge it in a data clean room with defined aggregates and cohorts for the customers it wants to target. Advertisers may then query matched data and run their own analyses to understand potential campaign reach.
Independent Data Clean Rooms
Separate from walled garden data clean rooms are independent data clean rooms. Independent third parties are not advertisers, but rather perform data matching across different data sets. Disney Select is a notable example of an independent data clean room, where a data manager is a conduit between the parties that want to share data.
An independent data clean room is where smaller companies lacking the customer data of a Walmart or an Amazon will band together to approximate larger data sets, sharing encrypted data to better understand how to interact with and advertise to a wider audience. Amazon itself recently validated this idea with the announcement it is extending its walled garden offering to allow independent groups to create and collaborate in clean rooms on AWS.
Data Quality Vigilance
Walled garden and independent data clean rooms share one important characteristic, which is that the records they produce by combining data sets are only as good as the data that is shared before it is encrypted.
That is, to trust the output, data quality processes and advanced identity resolution steps must be completed as part of the entry criteria. As the data clean rooms have evolved from a concept to reality, there is still a lingering misconception that identity resolution is the responsibility of yet another party in addition to the parties that are putting encrypted data into the data clean room.
While this may be true of the walled gardens, which in general do not count identity resolution as a core capability, any company using an independent data clean room should confirm which party is responsible for identity resolution. In addition, if identity resolution is outsourced to a third party, inquiries should also be made about what the process involves.
A CDP’s Place in a Data Clean Room
A data clean room that handles data quality at the time data is ingested will then ensure that the resulting matched records are accurate, reflecting the latest, most relevant information about a customer.
In essence, a data clean room should not divorce itself from the responsibility it shares with any robust customer data platform (CDP), which is to take a pool of data, perform data quality steps that include data cleansing, normalization, enrichment when appropriate, match/merge tasks and attaching a persistent ID to produce a pristine Golden Record.
But for a clean room, instead of storing the matched and identified customer records, once persistent IDs are attached, the data are anonymized, cohort and aggregate information is calculated, and results are put into the data clean room. For a brand-to-brand sharing use case, data quality steps are completed for data sets from each party sharing data, again with privacy-persevering transformations happening before the data is inserted into the data clean room. Essentially, the process is the same as an ordinary CDP process, except with two (or more) sets of data.
As a still-evolving technology, data clean room use cases are still being sorted out, in a sort of trial phase to determine how to enhance customer experience. What is clear, however, is that their popularity is due in part because they help monetize audiences, users’ CX or brand data while honoring the data value exchange and preserving privacy. Factoring in the benefits to both the brand and the end customer, the sharing of anonymized data in a data clean room has potential to be a win-win for all interested parties.